JupyterLab 4.2

This is the current default version of JupyterLab on Jupyter-JSC. In the next sections we will describe how you can modify it to fit your needs, and give a brief overview of the installed software.

Systems Available

JupyterLab 4.2 is available on these systems:

• JUWELS
• JEDI
• JURECA
• JUSUF
• DEEP
• JSC-Cloud

On JSC-Cloud only files in /home/jovyan are stored persistently. Everything else will be lost after a restart.

Pre-installed kernels

The kernels listed in this documentation may not always be up-to-date, as they can change periodically. For the current list of available kernels, please check the web service. The configuration files used to install these kernels are stored here in our GitHub repository.

• Bash
• Cling
• Julia
• LFortran
• Octave
• PyDeepLearning
• PyEarthSystem
• PyQuantum
• PyVisualization
• R
• Ruby

You can select them by navigating to the Kernels and Extensions tab on the left side of your configuration.

Kernel customization

These tutorials were created for HPC systems. Kernel creation on cloud resources (like JSC-Cloud) might require different steps. We may offer some documentation specifically for cloud resources in the future.

• Create kernel with virtualenv
• Create kernel with conda
• Create kernel with pyenv
• Create a containerized kernel with singularity
• Modify or extend a running kernel

Extensions

The extensions listed in this documentation may not always be up-to-date, as they can change periodically. For the current list of available extensions, please check the web service. The configuration files used to install these extensions are stored here in our GitHub repository.

• Jupyter AI
• Jupyter Archive
• Jupyter Bokeh
• Jupyter Collaboration
• Jupyter Resource Usage
• Jupyter Server Proxy
• Jupyter Slurm Provisioner
• JupyterLab Code Formatter
• JupyterLab favorites
• JupyterLab Git
• JupyterLab GitHub
• JupyterLab GitLab
• JupyterLab H5Web
• JupyterLab LaTeX
• JupyterLab nvdashboard
• JupyterLab Recents
• JupyterLab Sidecar
• JupyterLab Spellchecker
• JupyterLab Tour
• JupyterLab Trame Manager
• JupyterLab VariableInspector

Most extensions are always loaded. Others can be activated to your liking in the Kernels and Extensions tab in your JupyterLab configuration.

Install your own Extensions

One of the standout features of Jupyter-JSC is the flexibility it offers users to install their own extensions. This allows users to fully customize their environment according to their specific needs, rather than being limited to the pre-installed tools. With just three easy steps, users can add any extension that fits their workflow. The following example shows the installation of the JupyterView Extension.

The workflow outlined above has been tested specifically with prebuilt extension. Extensions that do not follow this standard might require extra steps for installation. For best results, request that the extension provider updates their software.

1. Install python package

test -f /tmp/custom/load_jupyter_version.sh && source /tmp/custom/load_jupyter_version.sh # Load the jupyter environment, only required on JSC-Cloud
pip install --user jupyterview # or the extension you would like to install

You can check the success with these commands:

test -f /tmp/custom/load_jupyter_version.sh && source /tmp/custom/load_jupyter_version.sh # Load the jupyter environment, only required on JSC-Cloud
jupyter labextension list

You will find the installed extension in this list. The extension might be listed with the hint "enabled X" instead of "enabled OK", that should be an issue while using the extension.

2. Stop JupyterLab

Browse to the home page and Stop your running JupyterLab.

3. Start JupyterLab

Wait until the JupyterLab is fully stopped. Then start it again.

You should now be able to use the installed extension. If something's not working as expected, you may find additional information in the JupyterLab logs at /tmp/custom/logs/stdout.

Uninstall your own Extensions

To remove the previously installed extension, you have to set PYTHONNOUSERSITE to 0. Otherwise pip will not look into the locally installed packages.

test -f /tmp/custom/load_jupyter_version.sh && source /tmp/custom/load_jupyter_version.sh # Load the jupyter environment, only required on JSC-Cloud
PYTHONNOUSERSITE=0 pip uninstall jupyterview

After a restart of your JupyterLab, the extension will be fully removed.

Proxies

• MatLab
• NEST Desktop
• RStudio
• Xpra